基于动态随机森林算法的铜浮选精矿品位预测

Grade prediction of copper flotation concentrate based on dynamic random forest algorithm

精矿品位作为反映产品质量的重要指标之一,一直是行业内关注的重点。采用机器学习里的动态随机森林算法对铜浮选过程中的铜精矿品位进行了模拟预测。利用国内某大型铜矿选矿厂浮选流程中泡沫图像分析仪获取的泡沫特征参数及其它相关数据分析了动态随机森林算法的预测效果,结果表明,动态随机森林算法能较准确地预测下一个测量周期的铜精矿品位,可用于指导作业者及时调整作业决策,确保最终精矿品位的稳定。

As one of the important indicators reflecting product quality,concentrate grade has always been the focus of attention in the industry.The dynamic random forest algorithm in machine learning is used to simulate and predict the copper concentrate grade in copper flotation process.Using the foam characteristic parameters and other relevant data obtained by the foam image analyzer in the flotation process of a large copper concentrator in China,the prediction effect of dynamic random forest algorithm is analyzed.The results show that the dynamic random forest algorithm can accurately predict the copper concentrate grade in the next measurement cycle,and can be used to guide operators to adjust their operation decisions in time to ensure the stability of the final concentrate grade.